llvm/lib/CodeGen/AsmPrinter/AsmPrinterInlineAsm.cpp

   1 //===-- AsmPrinterInlineAsm.cpp - AsmPrinter Inline Asm Handling ----------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 // This file implements the inline assembler pieces of the AsmPrinter class.
  10 //
  11 //===----------------------------------------------------------------------===//
  12
  13 #include "llvm/ADT/SmallString.h"
  14 #include "llvm/ADT/SmallVector.h"
  15 #include "llvm/ADT/StringExtras.h"
  16 #include "llvm/ADT/Twine.h"
  17 #include "llvm/CodeGen/AsmPrinter.h"
  18 #include "llvm/CodeGen/MachineBasicBlock.h"
  19 #include "llvm/CodeGen/MachineFunction.h"
  20 #include "llvm/CodeGen/MachineModuleInfo.h"
  21 #include "llvm/CodeGen/TargetRegisterInfo.h"
  22 #include "llvm/CodeGen/TargetSubtargetInfo.h"
  23 #include "llvm/IR/Constants.h"
  24 #include "llvm/IR/DataLayout.h"
  25 #include "llvm/IR/DiagnosticInfo.h"
  26 #include "llvm/IR/InlineAsm.h"
  27 #include "llvm/IR/LLVMContext.h"
  28 #include "llvm/IR/Module.h"
  29 #include "llvm/MC/MCAsmInfo.h"
  30 #include "llvm/MC/MCInstrInfo.h"
  31 #include "llvm/MC/MCParser/MCAsmLexer.h"
  32 #include "llvm/MC/MCParser/MCTargetAsmParser.h"
  33 #include "llvm/MC/MCStreamer.h"
  34 #include "llvm/MC/MCSymbol.h"
  35 #include "llvm/MC/TargetRegistry.h"
  36 #include "llvm/Support/ErrorHandling.h"
  37 #include "llvm/Support/MemoryBuffer.h"
  38 #include "llvm/Support/SourceMgr.h"
  39 #include "llvm/Support/raw_ostream.h"
  40 #include "llvm/Target/TargetMachine.h"
  41 using namespace llvm;
  42
  43 #define DEBUG_TYPE "asm-printer"
  44
  45 unsigned AsmPrinter::addInlineAsmDiagBuffer(StringRef AsmStr,
  46                                             const MDNode *LocMDNode) const {
  47   MCContext &Context = MMI->getContext();
  48   Context.initInlineSourceManager();
  49   SourceMgr &SrcMgr = *Context.getInlineSourceManager();
  50   std::vector<const MDNode *> &LocInfos = Context.getLocInfos();
  51
  52   std::unique_ptr<MemoryBuffer> Buffer;
  53   // The inline asm source manager will outlive AsmStr, so make a copy of the
  54   // string for SourceMgr to own.
  55   Buffer = MemoryBuffer::getMemBufferCopy(AsmStr, "<inline asm>");
  56
  57   // Tell SrcMgr about this buffer, it takes ownership of the buffer.
  58   unsigned BufNum = SrcMgr.AddNewSourceBuffer(std::move(Buffer), SMLoc());
  59
  60   // Store LocMDNode in DiagInfo, using BufNum as an identifier.
  61   if (LocMDNode) {
  62     LocInfos.resize(BufNum);
  63     LocInfos[BufNum - 1] = LocMDNode;
  64   }
  65
  66   return BufNum;
  67 }
  68
  69
  70 /// EmitInlineAsm - Emit a blob of inline asm to the output streamer.
  71 void AsmPrinter::emitInlineAsm(StringRef Str, const MCSubtargetInfo &STI,
  72                                const MCTargetOptions &MCOptions,
  73                                const MDNode *LocMDNode,
  74                                InlineAsm::AsmDialect Dialect) const {
  75   assert(!Str.empty() && "Can't emit empty inline asm block");
  76
  77   // Remember if the buffer is nul terminated or not so we can avoid a copy.
  78   bool isNullTerminated = Str.back() == 0;
  79   if (isNullTerminated)
  80     Str = Str.substr(0, Str.size()-1);
  81
  82   // If the output streamer does not have mature MC support or the integrated
  83   // assembler has been disabled or not required, just emit the blob textually.
  84   // Otherwise parse the asm and emit it via MC support.
  85   // This is useful in case the asm parser doesn't handle something but the
  86   // system assembler does.
  87   const MCAsmInfo *MCAI = TM.getMCAsmInfo();
  88   assert(MCAI && "No MCAsmInfo");
  89   if (!MCAI->useIntegratedAssembler() &&
  90       !MCAI->parseInlineAsmUsingAsmParser() &&
  91       !OutStreamer->isIntegratedAssemblerRequired()) {
  92     emitInlineAsmStart();
  93     OutStreamer->emitRawText(Str);
  94     emitInlineAsmEnd(STI, nullptr);
  95     return;
  96   }
  97
  98   unsigned BufNum = addInlineAsmDiagBuffer(Str, LocMDNode);
  99   SourceMgr &SrcMgr = *MMI->getContext().getInlineSourceManager();
 100   SrcMgr.setIncludeDirs(MCOptions.IASSearchPaths);
 101
 102   std::unique_ptr<MCAsmParser> Parser(
 103       createMCAsmParser(SrcMgr, OutContext, *OutStreamer, *MAI, BufNum));
 104
 105   // We create a new MCInstrInfo here since we might be at the module level
 106   // and not have a MachineFunction to initialize the TargetInstrInfo from and
 107   // we only need MCInstrInfo for asm parsing. We create one unconditionally
 108   // because it's not subtarget dependent.
 109   std::unique_ptr<MCInstrInfo> MII(TM.getTarget().createMCInstrInfo());
 110   assert(MII && "Failed to create instruction info");
 111   std::unique_ptr<MCTargetAsmParser> TAP(TM.getTarget().createMCAsmParser(
 112       STI, *Parser, *MII, MCOptions));
 113   if (!TAP)
 114     report_fatal_error("Inline asm not supported by this streamer because"
 115                        " we don't have an asm parser for this target\n");
 116
 117   // Respect inlineasm dialect on X86 targets only
 118   if (TM.getTargetTriple().isX86()) {
 119     Parser->setAssemblerDialect(Dialect);
 120     // Enable lexing Masm binary and hex integer literals in intel inline
 121     // assembly.
 122     if (Dialect == InlineAsm::AD_Intel)
 123       Parser->getLexer().setLexMasmIntegers(true);
 124   }
 125   Parser->setTargetParser(*TAP);
 126
 127   emitInlineAsmStart();
 128   // Don't implicitly switch to the text section before the asm.
 129   (void)Parser->Run(/*NoInitialTextSection*/ true,
 130                     /*NoFinalize*/ true);
 131   emitInlineAsmEnd(STI, &TAP->getSTI());
 132 }
 133
 134 static void EmitInlineAsmStr(const char *AsmStr, const MachineInstr *MI,
 135                              MachineModuleInfo *MMI, const MCAsmInfo *MAI,
 136                              AsmPrinter *AP, uint64_t LocCookie,
 137                              raw_ostream &OS) {
 138   bool InputIsIntelDialect = MI->getInlineAsmDialect() == InlineAsm::AD_Intel;
 139
 140   if (InputIsIntelDialect) {
 141     // Switch to the inline assembly variant.
 142     OS << "\t.intel_syntax\n\t";
 143   }
 144
 145   int CurVariant = -1; // The number of the {.|.|.} region we are in.
 146   const char *LastEmitted = AsmStr; // One past the last character emitted.
 147   unsigned NumOperands = MI->getNumOperands();
 148
 149   int AsmPrinterVariant;
 150   if (InputIsIntelDialect)
 151     AsmPrinterVariant = 1; // X86MCAsmInfo.cpp's AsmWriterFlavorTy::Intel.
 152   else
 153     AsmPrinterVariant = MMI->getTarget().unqualifiedInlineAsmVariant();
 154
 155   // FIXME: Should this happen for `asm inteldialect` as well?
 156   if (!InputIsIntelDialect && MAI->getEmitGNUAsmStartIndentationMarker())
 157     OS << '\t';
 158
 159   while (*LastEmitted) {
 160     switch (*LastEmitted) {
 161     default: {
 162       // Not a special case, emit the string section literally.
 163       const char *LiteralEnd = LastEmitted+1;
 164       while (*LiteralEnd && *LiteralEnd != '{' && *LiteralEnd != '|' &&
 165              *LiteralEnd != '}' && *LiteralEnd != '$' && *LiteralEnd != '\n')
 166         ++LiteralEnd;
 167       if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
 168         OS.write(LastEmitted, LiteralEnd - LastEmitted);
 169       LastEmitted = LiteralEnd;
 170       break;
 171     }
 172     case '\n':
 173       ++LastEmitted;   // Consume newline character.
 174       OS << '\n';      // Indent code with newline.
 175       break;
 176     case '$': {
 177       ++LastEmitted;   // Consume '$' character.
 178       bool Done = true;
 179
 180       // Handle escapes.
 181       switch (*LastEmitted) {
 182       default: Done = false; break;
 183       case '$':     // $$ -> $
 184         if (!InputIsIntelDialect)
 185           if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
 186             OS << '$';
 187         ++LastEmitted;  // Consume second '$' character.
 188         break;
 189       case '(':        // $( -> same as GCC's { character.
 190         ++LastEmitted; // Consume '(' character.
 191         if (CurVariant != -1)
 192           report_fatal_error("Nested variants found in inline asm string: '" +
 193                              Twine(AsmStr) + "'");
 194         CurVariant = 0; // We're in the first variant now.
 195         break;
 196       case '|':
 197         ++LastEmitted; // Consume '|' character.
 198         if (CurVariant == -1)
 199           OS << '|'; // This is gcc's behavior for | outside a variant.
 200         else
 201           ++CurVariant; // We're in the next variant.
 202         break;
 203       case ')':        // $) -> same as GCC's } char.
 204         ++LastEmitted; // Consume ')' character.
 205         if (CurVariant == -1)
 206           OS << '}'; // This is gcc's behavior for } outside a variant.
 207         else
 208           CurVariant = -1;
 209         break;
 210       }
 211       if (Done) break;
 212
 213       bool HasCurlyBraces = false;
 214       if (*LastEmitted == '{') {     // ${variable}
 215         ++LastEmitted;               // Consume '{' character.
 216         HasCurlyBraces = true;
 217       }
 218
 219       // If we have ${:foo}, then this is not a real operand reference, it is a
 220       // "magic" string reference, just like in .td files.  Arrange to call
 221       // PrintSpecial.
 222       if (HasCurlyBraces && *LastEmitted == ':') {
 223         ++LastEmitted;
 224         const char *StrStart = LastEmitted;
 225         const char *StrEnd = strchr(StrStart, '}');
 226         if (!StrEnd)
 227           report_fatal_error("Unterminated ${:foo} operand in inline asm"
 228                              " string: '" + Twine(AsmStr) + "'");
 229         if (CurVariant == -1 || CurVariant == AsmPrinterVariant)
 230           AP->PrintSpecial(MI, OS, StringRef(StrStart, StrEnd - StrStart));
 231         LastEmitted = StrEnd+1;
 232         break;
 233       }
 234
 235       const char *IDStart = LastEmitted;
 236       const char *IDEnd = IDStart;
 237       while (isDigit(*IDEnd))
 238         ++IDEnd;
 239
 240       unsigned Val;
 241       if (StringRef(IDStart, IDEnd-IDStart).getAsInteger(10, Val))
 242         report_fatal_error("Bad $ operand number in inline asm string: '" +
 243                            Twine(AsmStr) + "'");
 244       LastEmitted = IDEnd;
 245
 246       if (Val >= NumOperands - 1)
 247         report_fatal_error("Invalid $ operand number in inline asm string: '" +
 248                            Twine(AsmStr) + "'");
 249
 250       char Modifier[2] = { 0, 0 };
 251
 252       if (HasCurlyBraces) {
 253         // If we have curly braces, check for a modifier character.  This
 254         // supports syntax like ${0:u}, which correspond to "%u0" in GCC asm.
 255         if (*LastEmitted == ':') {
 256           ++LastEmitted;    // Consume ':' character.
 257           if (*LastEmitted == 0)
 258             report_fatal_error("Bad ${:} expression in inline asm string: '" +
 259                                Twine(AsmStr) + "'");
 260
 261           Modifier[0] = *LastEmitted;
 262           ++LastEmitted;    // Consume modifier character.
 263         }
 264
 265         if (*LastEmitted != '}')
 266           report_fatal_error("Bad ${} expression in inline asm string: '" +
 267                              Twine(AsmStr) + "'");
 268         ++LastEmitted;    // Consume '}' character.
 269       }
 270
 271       // Okay, we finally have a value number.  Ask the target to print this
 272       // operand!
 273       if (CurVariant == -1 || CurVariant == AsmPrinterVariant) {
 274         unsigned OpNo = InlineAsm::MIOp_FirstOperand;
 275
 276         bool Error = false;
 277
 278         // Scan to find the machine operand number for the operand.
 279         for (; Val; --Val) {
 280           if (OpNo >= MI->getNumOperands())
 281             break;
 282           const InlineAsm::Flag F(MI->getOperand(OpNo).getImm());
 283           OpNo += F.getNumOperandRegisters() + 1;
 284         }
 285
 286         // We may have a location metadata attached to the end of the
 287         // instruction, and at no point should see metadata at any
 288         // other point while processing. It's an error if so.
 289         if (OpNo >= MI->getNumOperands() || MI->getOperand(OpNo).isMetadata()) {
 290           Error = true;
 291         } else {
 292           const InlineAsm::Flag F(MI->getOperand(OpNo).getImm());
 293           ++OpNo; // Skip over the ID number.
 294
 295           // FIXME: Shouldn't arch-independent output template handling go into
 296           // PrintAsmOperand?
 297           // Labels are target independent.
 298           if (MI->getOperand(OpNo).isBlockAddress()) {
 299             const BlockAddress *BA = MI->getOperand(OpNo).getBlockAddress();
 300             MCSymbol *Sym = AP->GetBlockAddressSymbol(BA);
 301             Sym->print(OS, AP->MAI);
 302             MMI->getContext().registerInlineAsmLabel(Sym);
 303           } else if (MI->getOperand(OpNo).isMBB()) {
 304             const MCSymbol *Sym = MI->getOperand(OpNo).getMBB()->getSymbol();
 305             Sym->print(OS, AP->MAI);
 306           } else if (F.isMemKind()) {
 307             Error = AP->PrintAsmMemoryOperand(
 308                 MI, OpNo, Modifier[0] ? Modifier : nullptr, OS);
 309           } else {
 310             Error = AP->PrintAsmOperand(MI, OpNo,
 311                                         Modifier[0] ? Modifier : nullptr, OS);
 312           }
 313         }
 314         if (Error) {
 315           const Function &Fn = MI->getMF()->getFunction();
 316           Fn.getContext().diagnose(DiagnosticInfoInlineAsm(
 317               LocCookie,
 318               "invalid operand in inline asm: '" + Twine(AsmStr) + "'"));
 319         }
 320       }
 321       break;
 322     }
 323     }
 324   }
 325   if (InputIsIntelDialect)
 326     OS << "\n\t.att_syntax";
 327   OS << '\n' << (char)0;  // null terminate string.
 328 }
 329
 330 /// This method formats and emits the specified machine instruction that is an
 331 /// inline asm.
 332 void AsmPrinter::emitInlineAsm(const MachineInstr *MI) const {
 333   assert(MI->isInlineAsm() && "printInlineAsm only works on inline asms");
 334
 335   // Disassemble the AsmStr, printing out the literal pieces, the operands, etc.
 336   const char *AsmStr = MI->getOperand(0).getSymbolName();
 337
 338   // If this asmstr is empty, just print the #APP/#NOAPP markers.
 339   // These are useful to see where empty asm's wound up.
 340   if (AsmStr[0] == 0) {
 341     OutStreamer->emitRawComment(MAI->getInlineAsmStart());
 342     OutStreamer->emitRawComment(MAI->getInlineAsmEnd());
 343     return;
 344   }
 345
 346   // Emit the #APP start marker.  This has to happen even if verbose-asm isn't
 347   // enabled, so we use emitRawComment.
 348   OutStreamer->emitRawComment(MAI->getInlineAsmStart());
 349
 350   const MDNode *LocMD = MI->getLocCookieMD();
 351   uint64_t LocCookie =
 352       LocMD
 353           ? mdconst::extract<ConstantInt>(LocMD->getOperand(0))->getZExtValue()
 354           : 0;
 355
 356   // Emit the inline asm to a temporary string so we can emit it through
 357   // EmitInlineAsm.
 358   SmallString<256> StringData;
 359   raw_svector_ostream OS(StringData);
 360
 361   AsmPrinter *AP = const_cast<AsmPrinter*>(this);
 362   EmitInlineAsmStr(AsmStr, MI, MMI, MAI, AP, LocCookie, OS);
 363
 364   // Emit warnings if we use reserved registers on the clobber list, as
 365   // that might lead to undefined behaviour.
 366   SmallVector<Register, 8> RestrRegs;
 367   const TargetRegisterInfo *TRI = MF->getSubtarget().getRegisterInfo();
 368   // Start with the first operand descriptor, and iterate over them.
 369   for (unsigned I = InlineAsm::MIOp_FirstOperand, NumOps = MI->getNumOperands();
 370        I < NumOps; ++I) {
 371     const MachineOperand &MO = MI->getOperand(I);
 372     if (!MO.isImm())
 373       continue;
 374     const InlineAsm::Flag F(MO.getImm());
 375     if (F.isClobberKind()) {
 376       Register Reg = MI->getOperand(I + 1).getReg();
 377       if (!TRI->isAsmClobberable(*MF, Reg))
 378         RestrRegs.push_back(Reg);
 379     }
 380     // Skip to one before the next operand descriptor, if it exists.
 381     I += F.getNumOperandRegisters();
 382   }
 383
 384   if (!RestrRegs.empty()) {
 385     std::string Msg = "inline asm clobber list contains reserved registers: ";
 386     ListSeparator LS;
 387     for (const Register RR : RestrRegs) {
 388       Msg += LS;
 389       Msg += TRI->getRegAsmName(RR);
 390     }
 391
 392     const Function &Fn = MF->getFunction();
 393     const char *Note =
 394         "Reserved registers on the clobber list may not be "
 395         "preserved across the asm statement, and clobbering them may "
 396         "lead to undefined behaviour.";
 397     LLVMContext &Ctx = Fn.getContext();
 398     Ctx.diagnose(DiagnosticInfoInlineAsm(LocCookie, Msg,
 399                                          DiagnosticSeverity::DS_Warning));
 400     Ctx.diagnose(
 401         DiagnosticInfoInlineAsm(LocCookie, Note, DiagnosticSeverity::DS_Note));
 402
 403     for (const Register RR : RestrRegs) {
 404       if (std::optional<std::string> reason =
 405               TRI->explainReservedReg(*MF, RR)) {
 406         Ctx.diagnose(DiagnosticInfoInlineAsm(LocCookie, *reason,
 407                                              DiagnosticSeverity::DS_Note));
 408       }
 409     }
 410   }
 411
 412   emitInlineAsm(StringData, getSubtargetInfo(), TM.Options.MCOptions, LocMD,
 413                 MI->getInlineAsmDialect());
 414
 415   // Emit the #NOAPP end marker.  This has to happen even if verbose-asm isn't
 416   // enabled, so we use emitRawComment.
 417   OutStreamer->emitRawComment(MAI->getInlineAsmEnd());
 418 }
 419
 420 /// PrintSpecial - Print information related to the specified machine instr
 421 /// that is independent of the operand, and may be independent of the instr
 422 /// itself.  This can be useful for portably encoding the comment character
 423 /// or other bits of target-specific knowledge into the asmstrings.  The
 424 /// syntax used is ${:comment}.  Targets can override this to add support
 425 /// for their own strange codes.
 426 void AsmPrinter::PrintSpecial(const MachineInstr *MI, raw_ostream &OS,
 427                               StringRef Code) const {
 428   if (Code == "private") {
 429     const DataLayout &DL = MF->getDataLayout();
 430     OS << DL.getPrivateGlobalPrefix();
 431   } else if (Code == "comment") {
 432     OS << MAI->getCommentString();
 433   } else if (Code == "uid") {
 434     // Comparing the address of MI isn't sufficient, because machineinstrs may
 435     // be allocated to the same address across functions.
 436
 437     // If this is a new LastFn instruction, bump the counter.
 438     if (LastMI != MI || LastFn != getFunctionNumber()) {
 439       ++Counter;
 440       LastMI = MI;
 441       LastFn = getFunctionNumber();
 442     }
 443     OS << Counter;
 444   } else {
 445     std::string msg;
 446     raw_string_ostream Msg(msg);
 447     Msg << "Unknown special formatter '" << Code
 448          << "' for machine instr: " << *MI;
 449     report_fatal_error(Twine(Msg.str()));
 450   }
 451 }
 452
 453 void AsmPrinter::PrintSymbolOperand(const MachineOperand &MO, raw_ostream &OS) {
 454   assert(MO.isGlobal() && "caller should check MO.isGlobal");
 455   getSymbolPreferLocal(*MO.getGlobal())->print(OS, MAI);
 456   printOffset(MO.getOffset(), OS);
 457 }
 458
 459 /// PrintAsmOperand - Print the specified operand of MI, an INLINEASM
 460 /// instruction, using the specified assembler variant.  Targets should
 461 /// override this to format as appropriate for machine specific ExtraCodes
 462 /// or when the arch-independent handling would be too complex otherwise.
 463 bool AsmPrinter::PrintAsmOperand(const MachineInstr *MI, unsigned OpNo,
 464                                  const char *ExtraCode, raw_ostream &O) {
 465   // Does this asm operand have a single letter operand modifier?
 466   if (ExtraCode && ExtraCode[0]) {
 467     if (ExtraCode[1] != 0) return true; // Unknown modifier.
 468
 469     // https://gcc.gnu.org/onlinedocs/gccint/Output-Template.html
 470     const MachineOperand &MO = MI->getOperand(OpNo);
 471     switch (ExtraCode[0]) {
 472     default:
 473       return true;  // Unknown modifier.
 474     case 'a': // Print as memory address.
 475       if (MO.isReg()) {
 476         PrintAsmMemoryOperand(MI, OpNo, nullptr, O);
 477         return false;
 478       }
 479       [[fallthrough]]; // GCC allows '%a' to behave like '%c' with immediates.
 480     case 'c': // Substitute immediate value without immediate syntax
 481       if (MO.isImm()) {
 482         O << MO.getImm();
 483         return false;
 484       }
 485       if (MO.isGlobal()) {
 486         PrintSymbolOperand(MO, O);
 487         return false;
 488       }
 489       return true;
 490     case 'n':  // Negate the immediate constant.
 491       if (!MO.isImm())
 492         return true;
 493       O << -MO.getImm();
 494       return false;
 495     case 's':  // The GCC deprecated s modifier
 496       if (!MO.isImm())
 497         return true;
 498       O << ((32 - MO.getImm()) & 31);
 499       return false;
 500     }
 501   }
 502   return true;
 503 }
 504
 505 bool AsmPrinter::PrintAsmMemoryOperand(const MachineInstr *MI, unsigned OpNo,
 506                                        const char *ExtraCode, raw_ostream &O) {
 507   // Target doesn't support this yet!
 508   return true;
 509 }
 510
 511 void AsmPrinter::emitInlineAsmStart() const {}
 512
 513 void AsmPrinter::emitInlineAsmEnd(const MCSubtargetInfo &StartInfo,
 514                                   const MCSubtargetInfo *EndInfo) const {}